Intensive personnel attendance channel based on AI vision

Abstract

The invention provides a dense person attendance checking channel based on AI vision, relates to the field of attendance registration, and effectively promotes persons to spontaneously and orderly queue for attendance checking through the setting of a self-checking attendance checking channel and an attendance checking system and a specific use mode, so as to reduce queue jumping, collision and other behaviors caused by dense attendance checking persons and improve the attendance checking efficiency. Meanwhile, the enthusiasm of the personnel for actively dissuasing and resolving conflicts among others can be improved, the probability of further activating the conflicts of the personnel is reduced, and the safety and the attendance checking efficiency in the attendance checking process of dense personnel are improved, which is a first self-appointment means, and then the attendance checking punishment of delayed authentication is performed on the queue jumping personnel, so that the attendance checking efficiency is improved. By means of two self-appointment means, namely, a reward system giving priority attendance times to the whole staff, the occurrence of contradictions and conflicts in the attendance process can be further reduced, people to be cared and the staff with the priority attendance times can be prevented from queuing up in a pedestrian path and directly communicating with a gate, and the safety protection of the people to be cared is improved.

Description

Technical Field

[0001] The present invention relates to the field of attendance registration, and in particular to an attendance channel for densely populated personnel based on AI vision. Background Art

[0002] Attendance is a record of employees' working hours to determine whether they are late, leave early, or absent from work. Traditional attendance methods generally collect employees' fingerprints to record attendance. With the rapid development of AI vision technology, attendance methods based on AI vision have been applied in people's real lives.

[0003] For example, the specification of Chinese patent CN109741476B discloses an attendance system, which includes a server, an image acquisition device and at least one gate channel; the image acquisition device acquires the facial image of the personnel in real time and sends the facial image of the personnel to the server; the server receives the facial image of the personnel, verifies the facial image of the personnel according to the pre-stored employee information, and if the verification is successful, records the attendance information of the employee corresponding to the facial image of the personnel, determines the first target gate channel according to the preset channel determination method, and sends the first door opening instruction to the first target gate channel; the first target gate channel opens the gate upon receiving the first door opening instruction, and the image acquisition device can capture the facial image when the personnel enters or exits, so only one image acquisition device is needed to complete the acquisition of the facial image, which greatly reduces the cost. In addition, there is no need to output prompt information, and the personnel do not need to wait for the acquisition of the facial image, and the attendance record can be performed without the personnel's perception.

[0004] For example, the specification of Chinese patent CN118212704A discloses a method for on-site personnel attendance based on AI technology. A blockchain service center module, a monitoring service terminal module and an intelligent detection and judgment module are established in the central system. The blockchain service center module includes a blockchain data system submodule and a central processing submodule. The intelligent detection and judgment module includes an Internet connection terminal submodule and an intelligent maintenance monitoring submodule. The data captured through facial capture when going to and from get off work is connected with the labor real-name system to complete the attendance records. An alarm is issued for people who are not in the facial database when going to and from get off work, and the security guard goes there immediately, thereby improving the attendance efficiency of site personnel and solving the problem of congestion and queues at the entrance of the site.

[0005] The existing method of simply using AI vision to capture faces to achieve attendance has certain inaccuracies, especially in crowded environments, where there are certain missed detections. The method of using identity recognition devices to authenticate people one by one is prone to crowded queues and congestion. In this case, it is very easy to jump the queue, collide with others, etc., which poses a threat to the safety of attendance personnel. Therefore, there is an urgent need for an attendance channel that can restrain personnel behavior and improve attendance security. Summary of the invention

[0006] In view of the above-mentioned prior art, the technical problem to be solved by the present invention is that in a crowded attendance environment, conflict behaviors such as queue cutting and mutual collision are prone to occur, thereby posing a threat to the safety of the attendance personnel.

[0007] In order to solve the above problems, the present invention provides an AI vision-based intensive personnel attendance channel, including an attendance system and a self-checking attendance channel, the attendance system includes an attendance subsystem, a self-checking subsystem and a priority subsystem, the self-checking subsystem includes an air pressure monitoring module, a short shutdown module, a timing module and an alarm module, the priority subsystem includes a priority attendance module, a delayed attendance module and a special database, the special database stores information on people to be cared for and information on personnel with priority attendance times and delayed attendance times; The self-check attendance channel includes a top plate and a plurality of side plates fixedly connected at the lower end thereof, a pedestrian channel is formed between adjacent side plates, a gate is fixedly installed inside the pedestrian channel, a visual identifier is fixedly installed at the upper end of the gate, a sensing bag is connected to the ends of adjacent side plates close to each other, an air cavity is formed between the sensing bag and the side plate, an air pressure sensor located in the air cavity is connected to the end surface of the side plate, and the air pressure sensor is electrically connected to the air pressure monitoring module; A pair of side panels at the edge are both slidably connected with electric doors, the electric doors are located between the gate machine and the induction bag, and ends of the pair of side panels at the edge that are away from each other are both fixedly connected with visual identifiers 2.

[0008] Furthermore, the attendance subsystem includes an identity recognition module, a pass module and a personnel database. Visual recognizer 1 and visual recognizer 2 are electrically connected to the identity recognition module, the gate is electrically connected to the pass module, the pass module is connected to the short stop module, and the personnel database is used to store information of all personnel using the self-check attendance channel.

[0009] A dense personnel attendance channel based on AI vision, the use method of which includes the following steps: S1. People queue up along the pedestrian passage. When they arrive at the gate, they are identified by visual recognition device 1. People who need to be cared for and those with priority attendance times can choose to have their first identity recognized by visual recognition device 2. After successful recognition, the electric door opens automatically. People can directly reach the gate without queuing. Then they are identified twice by visual recognition device 1. After successful recognition by visual recognition device 1, the passage module controls the gate to open for people to pass through. S2. When a person cuts in line or collides with another person in the pedestrian passage, due to the size limit of the pedestrian passage, the person will collide with the sensor bag and cause a significant change in the air pressure sensor monitoring data. At this time, the short stop module starts, and the passage module is controlled to close the gate, suspending the passage of people for the next T seconds. S3. When there is no significant change in the air pressure sensor data for a second time within T seconds, the gate is restored to normal use. When the air pressure sensor data changes significantly again within T seconds, the gate is kept closed and the alarm module is activated to notify the management staff to handle the situation on site.

[0010] Furthermore, a camera is provided inside the pedestrian passage, and the camera is fixedly connected to the inner top of the top plate. The self-checking subsystem also includes an AI monitoring module electrically connected to the camera. The on-site processing in step S3 includes the following operations: Based on the image information collected by the camera and the testimony information from the surrounding people, the management personnel confirms the queue-jumping person A, and then enters the information of queue-jumping person A into the special database, and imposes a delayed attendance penalty on him. The operation of the delayed attendance penalty is as follows: when queue-jumping person A passes through the self-check attendance channel next time, he must complete the identity recognition operation on any two different visual identifiers, and after the identity recognition is completed on one visual identifier for the first time, the corresponding gate will not open. After the identity recognition is completed on another visual identifier for the second time, the corresponding gate will open to allow the person to pass.

[0011] Furthermore, the electric door is electrically connected to the priority attendance module, and the delayed attendance module is connected to the identity recognition module.

[0012] A dense personnel attendance channel based on AI vision, the use method of which also includes the following steps: when the short shutdown module is not started for N consecutive days, all personnel in the personnel database are granted the right to use priority attendance once.

[0013] Furthermore, the width between a pair of sensing bags in the initial inflation state is 60-100 cm.

[0014] Furthermore, the sensing bag includes a circular frame, the inner end surface of the circular frame is fixedly connected to the edge end of the airbag, the end of the circular frame away from the airbag is fixedly connected to a sealing gasket, and the circular frame and the side plate are detachably connected by multiple bolts.

[0015] Furthermore, a mounting groove is provided on the end surface of the side plate, and the air pressure sensor is bonded to the inside of the mounting groove by an adhesive.

[0016] To sum up, this application effectively encourages personnel to queue up in an orderly manner spontaneously through the setting of self-inspection attendance channels and attendance systems and specific usage methods, effectively reduces conflicts such as queue-jumping and collisions caused by dense attendance personnel, reduces the occurrence of personnel safety accidents, and at the same time can increase the enthusiasm of personnel to spontaneously dissuade and resolve conflicts between others, reduce the probability of further intensification of personnel conflicts, and further improve the safety and attendance efficiency of dense personnel attendance. This is the first self-appointment means, and through the delayed attendance punishment of double-channel authentication for queue-jumping personnel and the reward system of giving priority attendance times to all employees, the occurrence of personnel conflicts in the attendance process can be further reduced. This is the second and third self-appointment means. People who need care and people with priority attendance times can be exempted from queuing in the pedestrian passage and go directly to the gate, thereby improving the safety protection of the people who are cared for. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The device of the first and second embodiments of the present application is three-dimensional Figure 1 ; Figure 2 This is a system diagram of the first and second implementation modes of the present application; Figure 3 The device of the first and second embodiments of the present application is three-dimensional Figure 2 ; Figure 4 The device of the first and second embodiments of the present application is three-dimensional Figure 3 ; Figure 5 This is a three-dimensional diagram of the device when the sensing capsule is disassembled in the second embodiment of the present application; Figure 6 The device of the first and second embodiments of the present application is three-dimensional Figure 4 ; Figure 7 This is a schematic diagram of the top structure of the device in the first and second embodiments of the present application; Figure 8 This is a three-dimensional diagram of the sensing capsule in the second embodiment of the present application.

[0018] Description of the numbers in the figure: 1 top plate, 2 side plate, 201 mounting groove, 3 camera, 4 gate, 5 visual identifier 1, 6 sensor bag, 61 circular frame, 62 air bag, 63 sealing pad, 7 air pressure sensor, 8 electric door, 9 visual identifier 2. DETAILED DESCRIPTION

[0019] Two embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0020] The first implementation method: This invention provides an AI vision-based dense personnel attendance channel, which can be applied to company buildings, factories and other places with dense personnel attendance. Figure 1 and Figure 2 , including attendance system and self-check attendance channel. The attendance system includes attendance subsystem, self-check subsystem and priority subsystem.

[0021] The attendance subsystem includes identity recognition module, access module and personnel database; The self-check subsystem includes an air pressure monitoring module, a short stop module, a timing module, an AI monitoring module and an alarm module. The short stop module is connected to the passage module; The priority subsystem includes a priority attendance module, a delayed attendance module and a special database. The special database stores information on people who need care and information on people with priority attendance times and delayed attendance times.

[0022] The self-check attendance channel includes a top plate 1 and a plurality of side plates 2 fixedly connected at its lower end, a pedestrian passage is formed between adjacent side plates 2, a gate 4 is fixedly installed inside the pedestrian passage, a visual identifier 5 is fixedly installed at the upper end of the gate 4, the visual identifier 5 is electrically connected to the identity recognition module, the gate 4 is electrically connected to the passage module, the personnel database is used to store the information of all personnel using the self-check attendance channel, during the crowded attendance time period, the personnel should line up in the pedestrian passage in an orderly manner, arrive at the gate 4 in turn, and then perform identity recognition on the visual identifier 5, when the facial data collected by the visual identifier 5 corresponds to the facial information of a certain person in the personnel database, the identity recognition module sends a command to allow passage to the passage module, and the passage module controls the gate 4 to open for the personnel to pass and complete the attendance.

[0023] The ends of adjacent side panels 2 that are close to each other are connected with sensing bags 6, and a closed air cavity is formed between the sensing bags 6 and the side panels 2. The end faces of the side panels 2 are connected with air pressure sensors 7 located in the air cavity, and the air pressure sensors 7 are electrically connected to the air pressure monitoring module. In this embodiment, the sensing bags 6 and the air pressure sensors 7 are both fixedly connected to the side panels 2. A camera 3 is provided inside the pedestrian passage. The camera 3 is fixedly connected to the inner top of the top panel 1. The camera 3 is electrically connected to the AI ​​monitoring module. The camera 3 is used to automatically collect and record the facial and morphological information of the on-site attendance personnel to facilitate later inquiries.

[0024] An electric door 8 is slidably connected to a pair of side panels 2 at the edge, and the electric door 8 is located between the gate machine 4 and the sensor bag 6. A visual identifier 2 9 is fixedly connected to the ends of the pair of side panels 2 at the edge that are away from each other. The visual identifier 2 9 is also electrically connected to the identity recognition module. The electric door 8 is electrically connected to the priority attendance module. The visual identifier 2 9 can be used by people to be cared for and people with priority attendance times. The people to be cared for include pregnant women, people with limited mobility, etc. Since there are dense crowds of people during the attendance time period, in order to improve the safety of attendance of the people to be cared for, a priority attendance channel can be set for them. When the facial information collected by the visual identifier 2 9 corresponds to a certain facial information in the special database, the priority attendance module controls the electric door 8 to open for the person to pass. The person can directly reach the gate machine 4, and then perform identity recognition on the visual identifier 1 5, so that he does not need to queue in the pedestrian passage.

[0025] In the initial inflation state, the width between a pair of sensing bags 6 is 60-100cm. The passage within this size range can meet the normal walking needs of an adult, but it is difficult to meet the width for two people to pass (the normal width for two people to pass is generally 120cm and above). When a person cuts in line in the pedestrian passage, it is easy to cause some people (the queue cutters or the people who are queuing) to be squeezed and collided with the sensing bag 6 on one side, thereby causing the air pressure in the air cavity to change. Therefore, the monitoring data changes of the air pressure sensor 7 can effectively reflect whether there are behaviors such as attendance personnel cutting in line and conflicts.

[0026] A dense personnel attendance channel based on AI vision, the use method of which includes the following steps: S1. People line up along the pedestrian passage. When they arrive at the gate 4, they are identified by the visual identifier 1 5. People who are waiting for care and those with priority attendance times can choose to perform the first identity identification on the visual identifier 2 9. After successful identification, the electric door 8 automatically opens. People directly arrive at the gate 4 without queuing. Then they are identified twice by the visual identifier 1 5. After successful identification by the visual identifier 1 5, the passage module controls the gate 4 to open for people to pass; S2. When a person cuts in line or bumps into another person in the pedestrian passage, due to the size limit of the pedestrian passage, the cutting in line operation will cause a collision and squeezing between the person and the sensor bag 6, causing a significant change in the monitoring data of the air pressure sensor 7. At this time, the short stop module starts, controls the passage module to close the gate 4, and suspends the passage of people for the next T seconds; S3. When there is no significant change in the data of the air pressure sensor 7 within T seconds, the normal use of the gate 4 is restored. When there is significant data change in the air pressure sensor 7 again within T seconds, the gate 4 is kept closed and the alarm module is activated to notify the management personnel to handle the situation on site.

[0027] Steps S2 and S3 can effectively reduce conflict behaviors such as queue-jumping and collision, and can also increase the enthusiasm of people to spontaneously dissuade and resolve conflicts between others, as follows: First, when a person cuts in line, collides, etc., causing the gate machine 4 to be shut down for a short period of time, all people in the pedestrian passage will not be able to continue to pass through this gate machine 4, and the attendance progress will be delayed. Therefore, from the perspective of their own interests, this operation method makes it difficult for people to actively cut in line, conflict with others, etc. Second, when there is a situation where there is a conflict between people cutting in line and causing the gate machine 4 to be shut down for a short period of time, in order to resume the use of the gate machine 4 as soon as possible, the conflicting people will tend to take the initiative to reconcile with others, reduce the situation of intensified conflicts, and reduce the probability of colliding with the sensor capsule 6 again. At the same time, other people in the queue will tend to abandon the attitude of indifference, actively dissuade both parties, resolve the conflicts between the two parties, and resume the use of the gate machine 4 on time, so as to smoothly complete their own attendance process.

[0028] Therefore, this application can, to a certain extent, encourage personnel to queue up in an orderly manner spontaneously through the setting of self-inspection attendance channels, attendance system and specific usage methods, effectively reduce behaviors such as queue jumping and collisions caused by dense attendance personnel, and reduce the occurrence of personnel safety accidents. At the same time, it can also increase the enthusiasm of personnel to spontaneously dissuade and resolve behaviors such as queue jumping and conflicts, reduce the probability of further intensification of personnel conflicts, and further improve the safety and attendance efficiency in the process of dense attendance of personnel. This is the first means of self-regulation.

[0029] The on-site processing in step S3 includes the following operations: the manager confirms the queue-jumping person A based on the image information collected by the camera 3 and the evidence information of the surrounding persons, and then enters the information of the queue-jumping person A into the special database and imposes a late attendance penalty on him; The operation of delayed attendance penalty is as follows: when the queue-jumping person A passes through the self-check attendance channel next time, he needs to complete the identity recognition operation on any two different visual identifiers 5. The delayed attendance module is connected to the identity recognition module, which sends the personnel information with delayed attendance times in the special database to the identity recognition module, so that after the person completes the identity recognition on a visual identifier 5 for the first time, the identity recognition module does not send a command to allow passage to the passage module, that is, the corresponding gate 4 does not open. After the identity recognition is completed on another visual identifier 5 for the second time, the corresponding gate 4 opens to allow the person to pass.

[0030] By delaying the attendance penalty, people will have to queue up and undergo identity verification in two pedestrian lanes in turn during the next attendance process, which will greatly increase the time required for their next attendance. Therefore, considering their own interests and effectively avoiding delayed attendance penalties, people will be less likely to actively cut in line or collide with others, thereby further improving the safety and efficiency of the attendance process with a large number of people. This is the second means of self-regulation.

[0031] In addition, when the short shutdown module is not started for N consecutive days (that is, no collision between personnel occurs for N consecutive days), all personnel in the personnel database are granted the right to use priority attendance once. Personnel with priority attendance times can choose to directly use the visual identifier 2 9 for identity recognition at any time, and directly reach the corresponding gate 4 after passing the electric door 8, that is, they have obtained the right to be exempted from queuing in the pedestrian passage. Personnel can choose to use it when their attendance time is insufficient. Through this reward operation, the occurrence of behaviors such as queue jumping and collision during the attendance process can be further reduced. This is the third means of self-regulation.

[0032] Second implementation method: Based on the first embodiment, this embodiment makes the following specific settings for the sensing capsule 6 and the air pressure sensor 7: Figure 8 The sensing capsule 6 includes a circular frame 61, the inner end surface of the circular frame 61 is fixedly connected to the edge end of the airbag 62, and the end of the circular frame 61 away from the airbag 62 is fixedly connected to the sealing pad 63. Figure 5 As shown, the circular frame 61 and the side panel 2 are detachably connected by multiple bolts, an air cavity is formed between the airbag 62 and the side panel 2, and the sealing gasket 63 is made of elastic rubber material, which is in contact with the end surface of the side panel 2, which can improve the sealing between the circular frame 61 and the side panel 2, so that the gas in the air cavity is not easy to escape to the outside, and the air pressure in the air cavity is maintained stable. An inflation tube (not shown in the figure) extending to the inside of the air cavity is fixedly connected to the circular frame 61, which is convenient for filling a proper amount of gas into the air cavity after installation, so that the airbag 62 is in a moderate expansion state (expansion saturation is 60%-80%), and it is not easy to rupture due to human squeezing. A mounting groove 201 is opened on the end surface of the side panel 2, and the air pressure sensor 7 is bonded to the inside of the mounting groove 201 by an adhesive.

[0033] The air pressure sensor 7 is installed in a hidden manner inside the installation groove 201, which can effectively reduce the physical damage of the air pressure sensor 7 to the attendance personnel such as squeezing and collision. Through the setting of the above-mentioned structure, the sensing bag 6 and the air pressure sensor 7 are both connected to the side panel 2 in a detachable manner, instead of the fixed connection method in implementation mode 1. When the air bag 62 is damaged such as leakage or breakage, or the air pressure sensor 7 is not sensitive in monitoring, the two can be disassembled and replaced. In addition, when the present application is used in a non-dense area, without considering safety accidents such as queuing and queue jumping, the sensing bag 6 and the air pressure sensor 7 can be used without installing them. The present application can be used as an ordinary attendance device to reduce the consumption of equipment such as the sensing bag 6 and the air pressure sensor 7 and the cost of electricity.

[0034] In view of current practical needs, the above-mentioned implementation mode adopted in this application is not limited to the scope of protection. Various changes made within the knowledge scope of technical personnel in this field without departing from the concept of this application still fall within the scope of protection of the present invention.

Claims

1. A dense personnel attendance channel based on AI vision, characterized by: It includes an attendance system and a self-checking attendance channel. The attendance system includes an attendance subsystem, a self-checking subsystem and a priority subsystem. The self-checking subsystem includes an air pressure monitoring module, a short shutdown module, a timing module and an alarm module. The priority subsystem includes a priority attendance module, a delayed attendance module and a special database. The special database stores information on people to be cared for and information on people with priority attendance times and delayed attendance times. The self-check attendance passage comprises a top plate (1) and a plurality of side plates (2) fixedly connected at the lower ends thereof, a pedestrian passage is formed between adjacent side plates (2), a gate machine (4) is fixedly installed inside the pedestrian passage, a visual identifier (5) is fixedly installed at the upper end of the gate machine (4), a sensing bag (6) is connected to the ends of adjacent side plates (2) close to each other, an air cavity is formed between the sensing bag (6) and the side plate (2), an air pressure sensor (7) located in the air cavity is connected to the end surface of the side plate (2), and the air pressure sensor (7) is electrically connected to the air pressure monitoring module; An electric door (8) is slidably connected to a pair of side panels (2) at the edge, and the electric door (8) is located between the gate machine (4) and the sensor bag (6). A second visual identifier (9) is fixedly connected to the ends of the pair of side panels (2) at the edge that are away from each other.

2. According to claim 1, a dense personnel attendance channel based on AI vision is characterized by: The attendance subsystem comprises an identity recognition module, a pass module and a personnel database, the visual identifier 1 (5) and the visual identifier 2 (9) are both electrically connected to the identity recognition module, the gate (4) is electrically connected to the pass module, the pass module is connected to the short stop module, and the personnel database is used to store information of all personnel using the self-check attendance channel.

3. According to claim 2, a dense personnel attendance channel based on AI vision is characterized by: The method of use includes the following steps: S1. People queue up along the pedestrian passage. When they arrive at the gate (4), they are identified by the visual identifier 1 (5). The caring group and the people with priority attendance times are identified by the visual identifier 2 (9). After successful identification, the electric door (8) opens automatically. The people go directly to the gate (4) without queuing. Then, they are identified by the visual identifier 1 (5). After successful identification by the visual identifier 1 (5), the passage module controls the gate (4) to open and allow the people to pass through. S2. When a person cuts in line or collides with another person in the pedestrian passage, due to the size limit of the pedestrian passage, the cutting in line operation will cause a collision and squeezing between the person and the sensing bag (6), causing a significant change in the monitoring data of the air pressure sensor (7). At this time, the short stop module is activated, and the control passage module closes the gate (4), suspending the passage of people for the next T seconds; S3. When the data of the air pressure sensor (7) does not change significantly for the second time within T seconds, the normal use of the gate (4) is restored. When the data of the air pressure sensor (7) changes significantly again within T seconds, the gate (4) is kept closed and the alarm module is activated to notify the management personnel to handle the situation on site.

4. According to claim 3, a dense personnel attendance channel based on AI vision is characterized by: A camera (3) is provided inside the pedestrian passage, and the camera (3) is fixedly connected to the inner top of the top plate (1). The self-checking subsystem also includes an AI monitoring module electrically connected to the camera (3). The on-site processing in step S3 includes the following operations: The management personnel confirms the queue-jumping person A based on the image information collected by the camera (3) and the identification information of the surrounding persons, and then enters the information of the queue-jumping person A into the special database and imposes a delayed attendance penalty on him. The operation of the delayed attendance penalty is as follows: when the queue-jumping person A passes through the self-check attendance channel next time, he needs to complete the identity recognition operation on any two different visual identifiers (5), and after the identity recognition is completed on one visual identifier (5) for the first time, the corresponding gate (4) will not be opened, and after the identity recognition is completed on another visual identifier (5) for the second time, the corresponding gate (4) will be opened to allow the person to pass.

5. According to claim 2, a dense personnel attendance channel based on AI vision is characterized by: The electric door (8) is electrically connected to the priority attendance module, and the delayed attendance module is connected to the identity recognition module.

6. According to claim 3, a dense personnel attendance channel based on AI vision is characterized by: The method of using the system also includes the following steps: when the short shutdown module is not started for N consecutive days, all personnel in the personnel database are granted the right to use the priority attendance once.

7. According to claim 1, a dense personnel attendance channel based on AI vision is characterized by: In the initial inflation state, the width between the pair of sensing bags (6) is 60-100 cm.

8. The AI ​​vision-based dense personnel attendance channel according to claim 1 is characterized by: The sensing bag (6) comprises a circular frame (61), the inner end surface of the circular frame (61) is fixedly connected to the edge end of the air bag (62), the end of the circular frame (61) away from the air bag (62) is fixedly connected to a sealing pad (63), and the circular frame (61) is connected to the side plate (2) by a plurality of bolts.

9. The AI ​​vision-based dense personnel attendance channel according to claim 1 is characterized by: The end surface of the side plate (2) is provided with a mounting groove (201), and the air pressure sensor (7) is bonded to the inside of the mounting groove (201) by means of an adhesive.

Images